Latest / EHA Unplugged / The Current Diagnostic Approach to MDS (with Dr Onima Chowdhury)

Transcript

The Current Diagnostic Approach to MDS (with Dr Onima Chowdhury) · EHA Unplugged

0:07
Greetings and welcome to EHA. Unplugged, the official podcast
0:11
channel of the European Hematology Association.
0:14
EHA. So good morning.
0:17
Today I'm going to talk to only Chaudry.
0:21
She is hematologist from the University Hospital in Oxford
0:26
and she works both with the clinics and, and in research.
0:31
So a very warm welcome to Yoani. We're going to talk a little bit
0:35
about the diagnostic work up of MD's and, and what's needed and
0:40
maybe what we don't need and how we can do this best in different
0:44
situations. So, so maybe you could just
0:48
start by explaining us what you normally do or what do you think
0:52
it's important. Yeah, yeah.
0:54
Sure. So I guess thanks very much for
0:56
inviting me and the the basics are reasonably sort of well
1:01
established and straightforward. So to diagnose myelodysplastic
1:05
syndromes, you need cytopenia. So obviously a full blood count
1:08
is absolutely essential. And then following on from that,
1:11
morphology is really the mainstay of diagnosing
1:16
myelodysplastic syndrome. So typically that would be with
1:19
a bone marrow biopsy, but full blood count and a film can also
1:22
be very helpful. So normally a blood count and
1:26
then bone marrow cytomorphology would be the real basics.
1:30
And then moving on from that, there are, you know, more
1:34
detailed diagnostic tests that are required.
1:37
But I think I would like to talk a little bit about the
1:39
morphology of MD's because I think that's actually according
1:42
to what I, I as a clinician see. I sometimes think it's very
1:47
bearable how how how this is described.
1:49
I mean particularly at. Low risk MD's right?
1:53
I think that's really true and there are numerous sort of
1:55
International Studies that have been published saying that
1:57
there's a lot of inter observer variability in terms of what I
2:03
would call dysplasia and what someone else would call
2:05
dysplasia. So like you say, in the lower
2:07
risk of patients, it's a much bigger deal.
2:09
Or if you've got, you know, barn door blasts, really crazy sort
2:14
of red cell morphology, it's a lot more obvious.
2:16
No clear other reactive possible 'cause.
2:20
But obviously in the majority of our patients, that's not the
2:22
case. It's an older person who's gone
2:25
to their practitioner for, you know, a tiredness could be a
2:29
number of things in the background, mildly low blood
2:32
counts, a blood film that is potentially a little bit
2:36
abnormal. And typically we're supposed to
2:39
say more than 10% of the cells in a particular lineage are
2:42
dysplastic. Yeah.
2:44
So just to, just to, just to, to, to emphasize on that.
2:47
So it's any lineage right now. I mean, it doesn't have to be
2:50
several lineages, just any lineage that's just plastic.
2:53
Yes, exactly. And ideally more than 10% of the
2:56
cells. And again, The Who suggests that
2:58
you look at 500 cells ideally. And on a regular day, if I was
3:04
looking at a number of marrows, that's probably not the case for
3:06
most people. I think we'd all admit that.
3:09
So I think you're, you know, 200 is probably what most people
3:14
would be looking at, yeah. So, so if we go on from
3:19
morphology, what's, what's the next you would do?
3:22
Yeah. So I think we and I think
3:25
everyone really would try and get a trephine biopsy, which can
3:28
be very helpful in terms of not just the architecture but also
3:32
the cellularity is much more reliable on a trephine.
3:36
And also looking for additional features like fibrosis can be
3:39
helpful if you're thinking about an MD's or an MPN overlap,
3:42
that's often really useful I think to have a client.
3:45
Fibrosis is also important for prognostication, isn't it?
3:48
Yes. Exactly.
3:49
So your prognosis is not so good if you have fibrosis.
3:52
So that's very, very important. And also just looking for
3:56
additional features that may indicate there's a mast cell
3:59
disorder, you know, that has the dysplasia associated with it or
4:02
something along those lines. So I think a trefine certainly
4:06
diagnosis, not necessarily we might talk about following each
4:09
time you might do a marrow, but at the baseline trefine is very
4:13
helpful. And then I guess you move on to
4:16
cytogenetics. So typically a karyotype would
4:20
be of 20 cells and we would always request.
4:25
Sometimes that fails, which again we can talk about, but
4:28
ideally you'd be looking at 20 metaphase chromosomes to look
4:31
for and. Simply because cells will not
4:33
grow in the dish or. Yeah, exactly.
4:35
And you know, if we've got a poor quality sample, there's
4:38
been any delay if you're not looking at them locally, but
4:41
sending away it can be problematic.
4:44
But yeah, so that's and only 50% of patients with myelodysplastic
4:50
syndromes or neoplasms as we now call them, have a karyotypic
4:54
abnormality. But so that's really important.
4:57
And we also, I guess this varies probably from practice to
5:00
practice. We would always do a panel of
5:03
FISH probes, so fluorescence in situ hybridisation probes,
5:07
particularly if the karyotype has failed, looking for things
5:10
that are very typical of MTS, so chromosome 578 or P 53 related
5:17
abnormalities. So.
5:19
So you do that singly, or you only do it well if you don't, if
5:22
your cytogenetics fail. We typically only do it if our
5:25
cytogenetics fail or if there's a specific speed is often
5:31
another reason to do it. So the FISH will come back a lot
5:33
more quickly than your carrier type will do in most settings.
5:37
And in some, some cases we would also do a array if we have
5:42
failed carrier type. We're really worried looking for
5:44
a clonal marker. So that's a more widespread look
5:49
at the chromosome at the chromosome level for any
5:51
abnormalities that we might pick up.
5:52
But that that may not be available, I think to to a lot
5:56
of clinicians. No, I think you're right.
5:59
That is probably not what people usually do.
6:02
So, so in order to look for this Displacia, you know, do you use
6:05
any tools to is there any I think they can do online or
6:09
Yeah. We, we don't usually use it, but
6:12
there are definitely, certainly in the research sphere and
6:14
coming forward towards the clinic, there's a lot of
6:17
interest in using artificial intelligence and all its sorts
6:20
of guises. So I mean, I always view it that
6:22
certain things, you know, you really need an expert clinician
6:26
for sort of the nuanced diagnosis things where it's, you
6:29
know, there are unusual features, but actually we know
6:32
that computers are much better at picking up patterns than than
6:36
we are. And the reliability and the
6:37
reproducibility. So things are we know there's a
6:40
lot of inter observant variability.
6:42
So morphology, whether it be aspects or trefines, flow
6:45
cytometry is another one where actually you can use the machine
6:48
learning there. There's a lot of capacity for
6:53
improving on diagnostics and improving throughput in labs.
6:56
So actually we can get more patients diagnosed more
6:59
effectively. And I think that's definitely
7:02
something that we're looking at in terms of a research angle.
7:05
And there are numerous publications and things that
7:08
presented always at conferences in terms of how this could
7:11
really transform diagnostics. But it's still quite far from
7:15
the clinical. Yeah.
7:16
I think, I think it probably is in terms of validating these
7:20
things within a clinical trial setting.
7:23
But yeah, I think that's definitely the way that we
7:26
should be going. So we can focus more on the sort
7:29
of nuanced diagnostics rather and the the stuff that clearly
7:32
computers are better at than me. Which is actually very nice
7:37
because it also means that maybe more labs can come up with the
7:39
right diagnosis because often these patients ends up somewhere
7:43
else, at least in the beginning, right?
7:45
So. Completely.
7:46
So I guess systems vary in different places, but certainly
7:49
in the UK all hematopathology comes to certain centralised
7:55
centres, but that's only probably within the last, I
7:59
don't know, 10 years or so. So that because hematopathology
8:03
is a real skill. Yeah, it's really encouraging.
8:07
No, no, it's not easy and it is different.
8:10
Nothing to solid organ pathology.
8:13
So I think again, having that centralised system works, but
8:16
being able to do a lot of things locally would obviously help
8:20
throughput as well. And machine learning I think can
8:21
really help with that. Yeah, yeah.
8:24
Good. And what you, you mentioned flow
8:25
cytometry, is this necessary and what extra do we get out of flow
8:29
cytometry? Yeah.
8:30
So I guess for me with my interest in myelodysplastic
8:33
syndromes, I'm always, I'm always trying to think about how
8:35
we can incorporate it into our flow lab.
8:36
But it's not required. I think that's the bottom line
8:39
at the moment in terms of for diagnostics, it's not required
8:42
to to diagnose NDS, but there are again numerous publications
8:47
that they're scores that have been produced.
8:49
For example, example, the Agarta school, other schools where you
8:52
can use essentially ratios of different progenitor populations
8:56
with their myeloid or lymphoid. So simple parameters.
8:59
I think schools are always better if there are fewer inputs
9:03
that you can use to sort of say it is more likely that this
9:06
person will have MD's versus a reactive cause for their
9:09
dysplasia. And I guess the advantage of
9:11
flow is that it's quick. Again, you know, you can
9:14
hopefully get a result within 20 or 48 hours.
9:17
You're looking at live cells, you're looking at all of the
9:19
cells rather than already selecting off a population.
9:24
So I think there there is a place for it, but probably not
9:28
necessary in terms of routine diagnostics.
9:31
The other place that's clearly growing is in terms of looking
9:33
at sort of the immune landscape within.
9:36
Yeah. Exactly because that was what I
9:38
was thinking. We're always talking about how
9:40
important the immune system is for the development of MD's and
9:43
we sort of ignore this in a way. I mean, this is not normally
9:46
the, the, the things that we get out.
9:48
So. So how do you envision in the
9:49
future that this may be? I think I mean, again, it's
9:52
really important, isn't it? Because I mean, well, 25 to 30%
9:55
of patients with MD's have an inflammatory autoimmune
9:59
condition and this has been recognised for some time.
10:02
But I think it's becoming people are thinking more clearly that
10:04
are these linked. And so that's, that's important
10:07
from a clinical perspective. And then thinking about how some
10:10
of the treatments that we already have were, you know, the
10:12
only curative treatment is a bone marrow transplant, which is
10:14
effectively, you know, the ultimate immunotherapy.
10:18
Patients with low risk MD's or hypoplastic MD's that are
10:20
treated a bit more like aplastic anemia have had cyclistorian ATG
10:25
treatments like that, that clearly affect the immune
10:27
system. Azocytidine clearly has an
10:31
impact on the immune system. So there are lots of reasons to
10:34
think that the immune system is important and it clearly is.
10:37
And so actually having a panel that can incorporate immune
10:43
markers as well as sort of blast markers will be very helpful.
10:47
And, and as you know, there's a big effort to get this going
10:51
with the I4 MD's consortium. Exactly.
10:53
That's why I was. Trying to get there, yeah,
10:55
because. It's sincerely involved in that.
10:58
As you yeah, exactly. So that and that makes a lot of
11:01
sense. But I think the problem with
11:02
flow a bit like you said with morphology is there is a
11:06
relatively subjective objective element to I'm a big flow
11:09
cytometry fan. I've done loads of it and I
11:10
leave the flow cytometry service in our diagnostics lab.
11:13
But you need to overcome that in terms of making sure that people
11:17
are using the same markers. It's validated on big cohorts
11:22
because it's not always just sort of presence or absence.
11:24
It's it's there is a more nuance to it.
11:27
So I think having what you need is a big sort of international
11:30
working group working to get together for a clear panel of
11:34
markers that have clear clinical utility that a scientist in a
11:39
standard flow lab would be able to to work with.
11:42
Yeah. I think also this is one of the
11:44
aims actually of the. Yeah.
11:46
So moving on a little bit, what about, I mean, what about the
11:50
role of NGS in, in, in, in the diagnostic setting?
11:55
Yeah. So it's becoming increasingly
11:58
important, I think, and for a variety of reasons.
12:01
So I guess, well, it's a number of years now since we first
12:05
realised over 10 years that there are the recurrently
12:08
mutated genes in patients with myelodysplastic syndromes.
12:12
And unlike sort of the lucky people who look after MPNS,
12:15
they're not as clearly defined as they are with the
12:18
myeloproliferative neoplasms. But we have a collection of
12:21
mutations that are clearly important in terms of
12:24
identifying A clonal disorder, but not only that, they can help
12:29
with diagnosis. So in some settings with things
12:32
like P53 or SF3B1 would be the sort of the clear examples where
12:37
they have a clear prognostic and diagnostic relevance.
12:41
And now with the publication of the IPSS molecular score, we
12:47
have data now or the paper was published last year on 3000
12:52
patients, predominantly Europe, US and Japan, but relatively
12:56
international showing that you can incorporate mutations, 31
13:01
mutations in the paper and full blood count data and cytogenetic
13:05
data and be able to stratify patients more clearly using this
13:11
new score, which essentially over time will be incorporated
13:17
into clinical trials and is being incorporated into sort of
13:20
routine practice for us to be able to more accurately
13:23
prognosticate on our patients. So it's not really about
13:26
necessarily diagnosing them. These are all patients in the
13:29
paper that were had were being diagnosed with MD's.
13:33
But actually thinking about how you can offer a more accurate
13:36
prognosis and how that might impact on treatment and clinical
13:40
trial cohorts. I think it's clearly making a
13:42
huge difference, and not just in big labs that have been used to
13:45
doing it for years. I really agree on that.
13:49
I, I really think it helps a lot.
13:51
But, and, and I think maybe we should mention here also that
13:54
you can actually use it even though you don't have all the
13:56
genes, right, all the, the genes included in your panel.
13:59
But, but something that's very critical in this is, is actually
14:03
TPP 3 bialilic or more than allelic.
14:05
I mean, do you manage to get that out of your panel?
14:08
Yeah. So no, exactly.
14:11
I think it's. A.
14:14
Very common. Thing right?
14:16
Again, it's sort of honest. Well you know you have various
14:19
criteria that have been proposed in terms of how you might.
14:22
So the importance of this is obviously that again published a
14:26
few years ago I think 2020 that we all thought previously that
14:30
anyone with AP53 mutation was going to be absolutely dismal
14:34
prognosis. But what's come out of the work
14:37
really is that about 2/3 of the patients have a multi hit P53
14:40
mutation mutation with specifically with MD's.
14:44
Which means that they've got either 2 mutations, they might
14:48
have one mutation and then loss of 17 P or one mutation and a
14:54
copy neutral loss of heterozygosity on the other
14:58
chromosome. And so that would make it a
15:00
multi hit event. Whereas if you have a single hit
15:03
event, so just one mutation in a number of a number of different
15:07
mutations. Sometimes I think it's not clear
15:09
to other people that it's not a specific mutation, just one P53
15:13
mutation, that your prognosis in a number of studies now has been
15:17
shown to be nowhere near as poor as if you have a multi hit P53
15:22
mutation. But the problem in a diagnostics
15:24
lab is you're looking at bulk DNA.
15:27
So you're effectively taking your granulocytes or extracting
15:31
all your DNA and then sort of putting it all together and just
15:35
and coming up with a mutation. So I think if you've got 2
15:38
mutations that have got ideally some often similar VATS, then
15:43
you've got like double hits. If you've got AP 53 mutation,
15:48
your carrier type comes back and they've got loss of 17 P.
15:50
Then of course, yeah. Yeah, double hit.
15:52
The problems arrive when you start have trying to sort of
15:55
derive and I've cut offs ovarian allele frequencies.
15:58
And we also know talking a bit about research that if you look
16:01
at single cell data that it's not always brilliant.
16:04
So that if you think you know you've got AP 53 mutation of 20%
16:07
or 2030% or they'll have a multi hit mutation, they don't always
16:10
have a multi hit mutation when you look at single cell.
16:14
So it's not always clear cut. So we do need to be able to
16:17
improve on that. But broadly for a clinician
16:21
wondering whether their patient does have a poor prognosis P53
16:26
mutation, which is what it comes down to, I think in most
16:29
patients in a clinical setting, you you can unpick that from a
16:33
combination of diagnosis. Analysis.
16:36
Yes, that's true. Not just from the NGS.
16:39
The problem arises when you've got those sort of intermediate
16:42
variant allele frequency is in a patient who's quite well and
16:45
then am I going to tell them like is this?
16:47
Yeah. Yeah, exactly.
16:49
Or actually you're going to be one of those ones that does just
16:51
Bob along for ages because we definitely see those P53
16:53
mutations in various myeloid malignancies where actually they
16:58
just have a relatively low level of mutation for a very long time
17:02
and it doesn't any problems. And also I think there it's
17:05
very, I mean it's a challenge because when you discuss with
17:08
the transplanters, right, it's not really something that they,
17:12
they look into whether they are 353 limitations, right or not.
17:15
Right. Yeah.
17:16
OK to to move to another end of of of the spectrum.
17:21
What about when you have this patient that comes from the GDP
17:25
with a long lasting cytopenia? How important is the NTS then?
17:30
So if the patient older person with a long lasting cytopenia
17:35
with no other obvious 'cause I guess is what we're saying is
17:38
the GDP has been through routine bits.
17:42
So we kind of use an NGS panel in two different ways.
17:45
So if the if we think there's no way we're going to really want
17:49
to do a bone marrow biopsy in this patient for whatever
17:52
reason, other practical reasons, comorbidities, they're extremely
17:56
frail. The one hand, some of my
17:58
colleagues argue what's the point in doing a peripheral blab
18:01
NGS panel? You're not really going to do
18:03
very much. But I think often it is helpful
18:05
to give patients the diagnosis. It is helpful for the GPI agree
18:10
put an end to it and just say this may this is very likely to
18:15
be myelodysplastic syndrome. You know, the film is abnormal,
18:18
the counts fit, the phenotype fits and we can confirm that
18:22
with by showing, you know, a good level of a one or one or
18:27
more. Typically more than one mutation
18:29
would be enough to sort of this person's got monitors, plastic
18:32
syndrome and I would label them as MD's rather than getting too
18:35
much into the weeds of C Cus or I Cus, which I think it's really
18:38
difficult for non hematologists to really understand what you're
18:42
saying. So I think that that's one
18:46
reason we would often do it on the peripheral blood.
18:48
Now I appreciate it's not an issue.
18:49
So you would actually do it upfront.
18:50
We're trying to figure out how how we could do this more REST.
18:54
Yeah, I think we've gone almost too much the other way, to be
18:57
honest. So during COVID, when people
18:59
were not coming to see it at all, there was a massive spike
19:03
in peripheral blood, mild oil panels probably we need to come
19:06
down because the worry on the other hand is that you get a
19:09
mutation and you say, oh, this person's got MD's.
19:12
You then see the patient and realise that no, something else
19:14
is going on and they do need a bone marrow bopsy or do they do
19:17
need ACT because we're going to miss the fact that they've got
19:20
something else and people are. So I think it really takes an
19:24
expert to say that I'm confident that this is MD's and I'm just
19:29
using the myeloid panel in the blood to confirm my very high
19:33
level of suspicion rather than I'm using it as a blanket
19:37
screening tool to say anyone with this mutation has MD's
19:41
because I've definitely. So patients have got into
19:44
trouble I think with with that when it would've been too
19:46
reliant on it. So it's difficult.
19:48
So I think if they're going to have a bone marrow biopsy, just
19:51
do it on the bone marrow biopsy rather than, but if you're
19:54
thinking this patient probably isn't going to have one for
19:56
whatever reason, then I think it's not unreasonable to think
19:59
whether it'd be helpful. It's not.
20:01
But it is also helpful in a way because you can see if you have
20:03
an anemic elderly person, it might, they might actually
20:06
benefit from, from, from EPO and, and, and and transfusions
20:12
and things like that. So before you send them back to
20:14
the GP with Yeah. Yeah, Yeah, I think you're
20:17
right. I mean, I think that makes a big
20:19
difference. You would manage them the way
20:21
that you would a low risk MD's patient, but it's just given you
20:24
a high level of confidence. This is what's going on.
20:27
Yeah, yeah, yeah. So, so I think there's another
20:29
thing is that it's also nice for the maybe 2/3 that do not have a
20:32
mutation, because then you can say this is not what's going on.
20:35
I mean, it's a good discussion because I think you know, so,
20:38
so, so, so I, I want to ask you a question now because we've
20:41
been through the different ones. So so your hospital director
20:44
tells you that you need to cut the budget for diagnostics.
20:49
What would you cut for? I guess it, you know, I guess so
20:55
what's important and what's expensive?
20:57
I guess when you you get asked those questions by management,
21:00
so a lot of people probably say I'll just ditch morphology, but
21:03
I wouldn't ditch morphology because I think that it's still
21:05
really valuable and an essential part of diagnostic.
21:09
And of course, you might also find a lymphoma or.
21:12
Yeah, absolutely, definitely would, would, would keep keep
21:17
morphology. I think you would, you'd be hard
21:22
pushed to not keep a carrier type because again Dell 5QP53,
21:29
all these things that you can pick up, we have been picking up
21:31
using that. So I think if something had to
21:33
go, it probably, well, certainly the peripheral blood myeloid
21:37
panels that I'm just, I've just been talking about saying that
21:39
we would do, you could argue that actually we don't really
21:42
need to be doing them on the peripheral blood.
21:43
So for us that probably would save us quite a lot of money
21:46
because we do quite a few on the peripheral blood.
21:48
And then really think about who you're going to do them on when
21:50
you do a bone marrow biopsy. I think we'll probably be, so
21:54
you're almost going back to what we've done for years, you know,
21:56
and cytogenetics. I think that's probably still
22:00
the mainstay of, of, of diagnosis.
22:03
And then only in cases where you had a high index of suspicion,
22:07
but your carrier type had failed or it didn't show anything, then
22:10
you could sort of selectively we do an NGS panel and NGS panels
22:15
are are becoming cheaper. So we have an in house one which
22:18
just gets bigger and bigger and bigger because we're constantly
22:21
adding things in, often more for research purposes.
22:24
But you don't need that, you know the more.
22:26
How many? You only need this 30-40 teens.
22:29
Right. Yeah, exactly.
22:30
And and certainly for cloning hematopoiesis, there's talk
22:33
bringing in much smaller panels that.
22:35
Just heard about that. Sounds nice in a way about 18.
22:39
Yeah, exactly. Whereas because you know,
22:42
whereas you know in certain as we've got 40506070 genes, you
22:45
don't need that. So I think you could restrict
22:48
that in terms of the number of genes and the expense and also
22:51
the patients that you were offering it to.
22:54
That's what I would do. Yeah, OK.
22:57
Sounds good. So we talked about what we use
23:00
in the in the diagnostic situation, but what about follow
23:03
up? We've, we've tried to look at
23:05
this and we, we should probably publish this with getting some
23:08
other groups involved in that. What impact does a follow up NGS
23:13
panel really have because of these things are sort of
23:16
routinely requested and you can't well, at the moment ours,
23:23
most NGS panels, well, they're not, they're not validated for,
23:25
for sort of measurable residual disease monitoring.
23:29
So unlike, you know, real time PCR or you know, certain to play
23:34
cytometry assays that are validated for MRD monitoring in
23:38
AML. Sorry, I suppose we're talking
23:39
about. Yeah, that's.
23:40
The problem with that, with that, with MD's is obviously you
23:43
don't have the flow MRD and you don't have the the fusions that
23:48
often you would do for AML in terms of real time PCR.
23:51
So people often will then do an NGS panel thinking, oh, maybe
23:54
that will be helpful. It's helpful sometimes in terms
23:57
of clonal evolution, you know, if you're worried that something
24:00
else is going on. But in terms of we officially
24:04
only report mutations that are above 5% on the panel that we've
24:09
obviously when we then look at the files, we can then look in
24:12
more detail and see they've got lower levels of mutations.
24:14
But it shouldn't really be used to to influence patient
24:18
management at the moment. So I think we we need to be very
24:24
careful. So there isn't an obvious role
24:27
for follow up expensive myeloid panels, just routinely in
24:31
patients I think outside of a clinical trial setting.
24:35
I agree. Yeah, yeah.
24:37
I think it's. Important.
24:38
I can see a lot of it done. There's a lot of wasted on that,
24:41
in a way. Personnel in time, right?
24:44
Absolutely. In terms of resources and
24:46
reporting, you know, and it's, it all takes time.
24:49
As much as it seems quick, it's not that quick, you know?
24:52
Yeah. And one thing I think that might
24:55
be important to also talk about, what about the whole genome
24:58
sequencing? Yeah.
25:00
Does it have a place here? Can it replace what we're doing
25:04
on our? Yeah, yeah.
25:06
So I think whole genome sequencing again, I am often
25:12
sort of a step ahead in some ways.
25:14
It has certainly been. There's a New England Journal
25:16
paper a couple of years ago looking at adding in whole
25:21
genome sequencing to your routine pipeline for AML, saying
25:24
that it was feasible. But obviously this is that, you
25:26
know, big institutes that are used to running these.
25:29
Pipeline, I'm thinking ahead a little.
25:32
Bit, yeah, exactly. So I think the there is a big
25:35
push for whole genome sequencing.
25:38
Advantages are that you can get your copy number alterations as
25:43
well as your point mutations which is great.
25:47
So you can either use it in a targeted way or in a discovery
25:50
way as well as you get far more information.
25:54
The, the the downsides at the moment are one, the coverage is
25:58
not so good. So when you when we talk about
26:00
coverage you talk you what we mean is that when you sequence a
26:03
certain area, the number of reads you have or sort of the
26:07
number of times that bit of the genome is sequenced is far
26:11
shallower withholds you don't sequencing for an predominantly
26:14
from an expense perspective, but it's getting cheaper so that
26:18
they will be able. To, but that also means that you
26:20
won't won't find the small smaller clones in them, yeah?
26:24
Again, again, when people talk about P53, you know, you can
26:27
pick up whether it's multi hit, but you're still only looking at
26:31
vast. So you know, 20% and above, you
26:33
know, maybe a bit lower if you're lucky.
26:34
So no, you're absolutely right. So the small clones you can't
26:36
pick up on whole genome sequencing.
26:38
It's more expensive. The bioinformatic pipelines
26:42
aren't really not in place for it to be a quick sort of
26:45
throughput in a routine, routine lab that that needs more work.
26:50
But yes, you do get more information in that one test.
26:56
So I think that there is a role for it and we use it in patients
27:01
more where the diagnosis is unclear.
27:03
So not so much so into, again, as a discovery tool.
27:06
And sorry, the other thing that most whole genome sequencing,
27:09
bioinformatic hyperplanes require is the germline control.
27:12
So you can do tumour only, but most of them require a germline
27:16
control. And there's a lot of controversy
27:18
about what you should or shouldn't.
27:19
Or practically, you can use as a germline control to show that
27:23
these are acquired somatic mutations, which is not a
27:27
problem that you have with targeted panels.
27:31
No, of course not. And that takes me actually to to
27:34
germline disorders who also leads to MD's because how do you
27:38
pick up those? Yeah.
27:40
So we, I guess you go back in terms of clinical practice, you
27:44
think about patients and often we think about them as having to
27:48
be young. And then I guess certainly new
27:50
mutations that we've more recently learnt about like DDX
27:53
41 have shown us. What we always kind of should
27:56
have known is that this is not the case.
27:58
You know, you, you still, even if someone's often maybe would
28:02
have said in the 20s, thirties, 40s, or maybe there's something
28:04
germline going on, but you can be much older and they still
28:08
have a germline predisposition. So that's taught us a lot, I
28:10
think as adult hematologists, whereas the pediatricians are
28:13
obviously looking for it. But clearly things that would
28:15
point you in that direction or a family history or or a physical
28:18
phenotype that might make you think along those lines.
28:22
But, but germline predisposition is very important not only for
28:27
the individual and for the family screening, which is I
28:32
think it's what we found it very difficult in terms of working at
28:34
what to do with DDX 41 and certainly, yeah, and
28:40
particularly with regards to the family screening.
28:43
So which we need sort of more international clarity on and
28:47
really what we should be doing in terms of for the screening
28:50
for the patient. Obviously a lot of these for the
28:54
patients with, with MD's often you typically would have a
28:57
germline mutation. Often they may then acquire an
29:00
additional typical somatic mutation.
29:04
So the the management of those patients is somewhat clearer
29:08
than the family screening. The other thing I think germline
29:10
screening is really important to think about is that when you're
29:12
thinking about transplantation. So you've got a family history
29:16
or any indication that there may be a germline predisposition and
29:19
you're thinking about sibling donors or, or now we're doing a
29:21
lot more hapless donors. It's a major issue.
29:26
So not just for the patient, but for family members and for
29:30
future transplantation. And sorry, back to your original
29:33
question, how you pick them up on GS panels now of most people,
29:37
so probably would have incorporated DDX 41
29:39
specifically. But when you see other mutations
29:42
that may be more typical for a germ line mutation.
29:45
So things like, I don't know CBPR 4 rungs or things like that
29:49
where if you, if you've got a variant allele frequency that is
29:52
more typical for a germ line mutation, we would then ask for
29:57
just skin biopsy culture fibroblast and sequence that and
30:01
then send them for fat for genetic counselling if we we
30:08
have. And that is always the first
30:09
step, right? The genetic counselling.
30:11
Yeah, yeah. So that's not something we take
30:13
on at all in our practice hematologist.
30:16
That's something that's very special.
30:19
So what about the whole genome sequencing in that respect?
30:22
How how efficient is it to pick up the the germline?
30:26
Yeah. So I think personally in our
30:31
lab, I don't have as much to do with whole genome secrecy, but
30:33
it's it is, but it's more robust in terms of picking up germline
30:38
variants. For sure.
30:41
Yeah, good. So now we don't have no loan
30:44
interpreters for 10 years maybe. And we are getting a little bit
30:47
wiser every day I guess because there's so many publications.
30:50
But but how do you, what do you think about chip and that what
30:54
do you do if you pick up a mutation that is potentially
30:57
chip in your airport? Yeah.
30:58
So I think that that the first thing you've mentioned about
31:00
picking up a mutation is really important because there's a big
31:03
difference in, in what the huge sort of population studies of
31:08
200,000 in UK by a bank, other studies in the US, other
31:11
international collaborations are picking up.
31:13
And what we pick up in a diagnostics lab is totally
31:16
different because anyone who's come to a diagnostics lab has a
31:20
in general has a problem. They're not totally healthy.
31:24
So we pick up very little chip. So the chip, so for you to get a
31:30
myeloid panel, there is some filtering.
31:33
You know when I was saying earlier we do loads, we don't
31:35
do, we're not crazy, we don't do the one absolutely everyone.
31:38
But still you would generally would have had an abnormal full
31:40
blood count. So the ones that we then call
31:43
chip are the ones that then there's their counts then
31:45
recover and it was clearly a reactive course and then you
31:48
pick you've picked up mutation, but they've got normal counts.
31:52
All the other ones are the ones that come in query MPN Yeah,
31:55
well, actually so so I think, but it's to be honest, it's a
31:58
handful. Most of this is very much in the
32:02
research domain when you think about real chip and even when
32:06
you look at some of the biggest studies, they're not very clear
32:10
as to whether they're these these patients do have something
32:13
else going on and what their blood counts really are.
32:15
So are they? Because.
32:17
They're indeterminate potential because if you've got you know
32:22
by definition to be Chip, you have a somatic mutation more
32:27
than 2% variant allele frequency in a gene that is known to
32:31
predispose you to a blood cancer.
32:33
But some of these patients or and with normal blood count,
32:36
sorry, but so I think it's a very different what we see
32:40
really in a diagnostics lab compared to what sort of the
32:44
population wide studies are demonstrating.
32:47
So we certainly don't have yet a chip clinic as such.
32:53
And I think it's really about the, the reason to do it is to
32:58
sort of have prospective cohorts of patients going forward, not
33:03
because it's really going to make that much difference to the
33:06
majority of patients. Remembering that, you know, we
33:09
talk about blood cancer risk or even cardiovascular risk and
33:12
overall mortality, but you know, 95% of these patients are going
33:16
to be with real CHIP are going to be absolutely fine.
33:20
You know, there's a 1% risk of per year we think Max of
33:25
developing a blood cancer, a higher risk of cardiovascular
33:28
disease. So I think we need to be really
33:32
careful about how we, you know, run this into routine practice
33:36
research. Brilliant.
33:37
I mean, we should definitely be doing as much research as we can
33:40
on this, but actually to to it's hard to talk to patients I think
33:46
as well about, you know what this.
33:47
Yeah, because you're actually making healthy people sick by
33:50
telling them they have a. Majority exactly and not much
33:52
has been published on that. So there's there's a study of
33:56
patients in the US looking at whether patients wanted to know
33:59
about CHIP. It was a study of women who'd
34:01
had breast cancer. That's a very, very different
34:04
cohort of people. Very different.
34:05
It's very different. I mean there.
34:06
Are more path cancers? But I think it's very different
34:09
because I mean studies also in lymphoma, for example, showing
34:13
that it actually is associated with a poor prognosis.
34:16
Yeah, it absolutely at least. Yeah, yeah.
34:19
And then you can't, we shouldn't really be calling those pay
34:21
people Chip. And it's not of indeterminate
34:24
potential. They've declared themselves as
34:26
having potential. Having something, yeah.
34:30
You're right. So I think it's a good, but I
34:33
think potentially more from the cardiovascular side and
34:35
incorporating more of our cardiovascular colleagues into
34:37
these discussions or other inflammatory disorders borders
34:41
is more valuable than worrying about whether they're going to
34:43
get AML. Not having said that, obviously
34:46
couple of papers have now been published from the Cambridge
34:48
group and from Boston providing school, you know, school how you
34:52
might assess whether someone's high or low risk.
34:55
So incorporating full blood count data, which is really
34:58
interesting from a diagnostic perspective.
34:59
So MCV and RDW, whether there are other factors in your raw
35:05
full blood count data there probably are which can determine
35:09
whether you're more likely to develop cancer in the future.
35:13
AI would be very useful for. So whether you can before
35:17
patients you know earlier along the pathway, can you detect
35:20
these patients Obviously a very good question but without
35:22
alarming you know. Everyone, yeah.
35:25
So if you can identify like a high risk, high risk too or
35:29
yeah. Yeah.
35:30
And I think that is that is worth doing.
35:33
Yeah, very nice. I think this was very, I think
35:37
we came around everything. Now I'm I'm very happy.
35:40
Hope you're happy too. That was great, really enjoyed
35:42
it. And that the Education Committee
35:44
will be happy. So thank you very much for.
35:46
No worries. Thank you very much.